Canine blood oxygenation using ultra-low blood flow in an in vitro, single-circuit, extracorporeal membrane oxygenator model

Objective To ascertain the efficacy of a single circuit, in vitro extracorporeal membrane model at oxygenating blood and removing carbon dioxide (CO2) with ultra-low blood flow, in replicates either exposed to room air (fraction inspired oxygen 0.21) (n = 6) or ultra-low 100% oxygen flow (n = 6) driven by a linear peristaltic volumetric infusion pump. The effects of both replicates on free plasma haemoglobin (fHb) concentration was also determined. Furthermore, the ability of the oxygen replicates to maintain global oxygen delivery (DO2) was predicted in a theoretical model. Study Design In vitro, experimental study. Methods Twelve units of canine stored whole blood were used, with a median (minimum - maximum) volume of 465 (440 – 500) mL, packed cell volume of 0.5 (0.36 – 0.56) L L-1 and storage time of 3 (1 – 41) days. The blood circuit was constructed in the following order, assembled in series; blood reservoir, oxygenator, linear peristaltic infusion pump and tubing to complete the circuit by returning the blood to the reservoir. The water circuit was constructed by connecting a bath warmed to 44°C containing lactated Ringer’s solution to the water inlet of the oxygenator by tubing and a linear peristaltic infusion pump. Tubing connected to the water outlet of the oxygenator completed the water circuit by returning the water to the bath. Gas supply to the room air replicates was ensured by removing the gas inlet safety protection cap during assembly. For the oxygen replicates, an oxygen supply rig was constructed to split the oxygen flow to supply an ultra-low flow of oxygen (0.03 L minute-1) to each oxygen replicate. Before the sampling commenced, the blood and water phases of the oxygenators were primed. After 2 minutes of the blood and water circulation (both flows at 0.02 L minute-1), the first samples were collected (T0). Samples were collected for blood gas analysis post oxygenator (PaO2 and PaCO2) and pre-oxygenator (PvO2 and PvCO2). In the room air replicates, samples were collected hourly for the first 8 hours (T1, T2, T3, T4, T5, T6, T7 and T8) then at 24, 32, 48 and 56 hours (T24, T32, T48 and T56). In the oxygen replicates, the oxygen supply was connected after T0 and samples were then collected at 15-minute intervals for the first hour (T0.25, T0.5, T0.75 and T1) and then hourly for 8 hours (T2, T3, T4, T5, T6, T7 and T8) and then at 24 hours (T24). All the post-oxygenator samples in both replicates were centrifuged and analysed for fHb concentrations, except at T0.25, T0.5 and T0.75 in the oxygen replicates. Data was compared using a linear mixed model (fixed effect: time; random effect: replicates) and post-hoc analysis using Dunnet’s method within each replicate where each time point was compared to T0. Statistical significance was set at p < 0.05. A theoretical model predicting the effect on DO2 over a range of weights was constructed assuming that the oxygenator was augmenting mixed venous oxygen content. The effects on DO2 were extrapolated using PaO2, arterial oxyhaemoglobin saturation (SaO2) and haemoglobin concentrations (Hb) from the study as well as 2 hypoxaemia scenarios: hypoxic hypoxaemia (PaO2 40 mmHg; SaO2 75%) and anaemic hypoxaemia (Hb 6.7 g dL-1). These theoretical DO2 values were compared to a critical DO2 of dogs which is reported to be 9.8 mL kg-1 minute-1. Results All replicates were operational for the duration of the study period, except 2 of the room air replicates which failed due to thrombosis between T32 and T48. In the room air replicates, the PaO2 significantly increased from T0 during T1 to T8; the PaCO2 significantly decreased from T0, during T2 to T56. In the oxygen replicates, the PaO2 significantly increased from T0 for the entire study duration; the PaCO2 significantly decreased from T0 at all time points. In the room air replicates, the rate of change of fHb concentration did not change from T0 for the study duration. However, in the oxygen replicates, the rate of change of fHb concentration was increased from T0 at T3, T4 and T6. In the theoretical model, the predicted DO2 was maintained above the critical DO2 when calculated from study variables and the hypoxic hypoxaemia scenario. However, in the anaemic scenario, the predicted DO2 fell below the critical DO2. Conclusion and clinical relevance The extracorporeal membrane oxygenator configured for ultra-low blood and oxygen flow significantly increased the PaO2 and decreased the PaCO2 for 24 hours. Furthermore, the rate of change of fHb concentrations within the replicates indicate acceptable blood handling characteristics by the circuit components. An increase in DO2 was identified using the theoretical model and may clinically improve myocardial oxygenation in pathological conditions characterised by an oxygen debt. Further studies are needed to ascertain whether this study can be translated into clinical setting.